Skip to content

Recreate photos in the style of famous artists with neural networks

License

Notifications You must be signed in to change notification settings

jayelm/neural-art

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

25 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Not another neural style implementation!

Recreate photos in in the style of famous artists with machine learning. CSCI3341 Artificial Intelligence Final Project.

Usage

Make sure to download models with the util/download_models.sh script first. Thanks @jcjohnson!

python art.py --help

Examples

Create a (low quality) version of the image below. The first positional argument is the content image and the image(s) after are the style images. The further options instruct the algorithm to run for 200 iterations and output an image with a 256px maximum width, which should be doable relatively quickly on a personal computer.

python art.py examples/gasson.jpg examples/starry.jpg -n 200 -w 256

Further options include using the entire oeuvre of an artist as a stylistic target (which never works well), scaling the stylistic features by a factor, specifying whether to initialize the candidate image from random noise or the content image, and more. See the script's help for details.

The best Gassongram ever:

gasson starry gasson_final

Where do I get all this art?

util/wikiart-scraper.py contains a neat script which will automatically scrape every painting for a given artist from wikiart. Simply specify the URL component of the artist's name (e.g. pablo-picasso) in the ARTISTS array in the script.

About

Recreate photos in the style of famous artists with neural networks

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published